Vertical Accuracy Assessment of Freely Available Digital Elevation Models: Implications for Low-Relief Landscapes.

Global Digital Elevation Models (DEMs) are widely used in many developing countries to help mapping soil-scapes, and users rely on metadata on vertical accuracy (VA) given by product developers. VA is important in modelling topographic attributes useful for mapping and predicting soil-scapes. This paper summarizes a comparative analysis of vertical accuracy of eight freely available DEMs derived from EO datasets in a low-relief landscape of Argentina (Entre Rios province), where the main objective is to map complex soil cover patterns. These are: MERIT-DEM; SRTMGL1 v3; ASTGTM v3; AW3D30 v3.1; Copernicus GLO-30; NASA DEM HGT v1; SRTMGL1up-sampling; Argentina's MDE-Ar v2. Results show the Copernicus GLO-30 dataset was consistent with its nominal quality specifications, whereas the ASTGTM v3 from NASA/Jaxa ranked last. While these results are restricted to the geography of this study area, they reaffirm the need to assess DEM's quality independently of that provided by the product developers.